Users are provided with an ever increasing range of operations to interact with images in a user interface. As part of this interaction, a user may desire to select an object in the image, such as to change characteristics of the object using one or more editing objects, remove the object from the image, copy the image for inclusion as part of another image, and so forth.
Conventional techniques used to select an object within an image, however, typically required a user to manually trace a border of the object being selected. Although this works well for objects having regular uniform boundaries, such as those found in simple shapes, these conventional techniques do not work well for objects having irregular shapes. Accordingly, conventional techniques could involve detailed and frustrating tracing of irregular borders which could be inefficient and result in significant inaccuracies.